---

Similar Listener Networks

Methodology

Similar Listener Networks analyze listener's preferences in music, then connect listeners to others with similar tastes and compare favorite tracks, albums, and artists in order to recommend new music to the listener. Similarity can be determined in a number of different ways and a statistical rating of similarity will typically be assigned. This method can be used in conjunction with similar music networks, as Last.fm illustrates, to cyclically recommend other listeners based on similar music and recommend music based on similar listeners. Similar Listener Networks can also operate either indirectly or directly - tracks can be recommended based on the combined preferences of many similar listeners or directly from one listener to another, via reviews on forums or blogs. As is the case with Last.fm, Similar Listener Networks can take on both forms simultaneously, where listeners receive both a general indirect list of recommended artists based on the preferences of similar listeners, and can also interact socially with other similar listeners and recommend music directly.

Pros and Cons

Similar Listener Networks are ideal for connecting listeners to others with shared music interests. Musical understanding is relative to the listener, and therefore it is logical that finding listeners with similar taste in, and thus similar understanding of, music would be the most accurate way to recommend music to listeners. However, there are a number of issues that arise in similar listener network systems. The most important is the cyclical nature of these recommendation systems. Because people are recommended more music that they are likely to enjoy, they become further and further entrenched in their preferences in music and aren't given the opportunity to expand their tastes in a manner appropriate to them.